Blog post · 2025
GDPval: measuring the performance of our models on real-world tasks
| URL | https://openai.com/index/gdpval/ |
|---|---|
| Topics | ReplicHAI / Agent Evals / 07 General, long-horizon & economic tasks |
| Tags | agent-eval |
Abstract
Measures: Occupational tasks graded by experts. Metric: Win/tie rate.
BibTeX
@misc{openai_gdpval_2025,
title = {{GDPval}: measuring the performance of our models on real-world tasks},
url = {https://openai.com/index/gdpval/},
abstract = {Measures: Occupational tasks graded by experts. Metric: Win/tie rate.},
author = {{OpenAI}},
month = sep,
year = {2025},
keywords = {agent-eval},
}